The Rise of Algorithmic Policing in the Federal Government
Law enforcement agencies across the United States have increasingly adopted different types of quasi-“precrime” monitoring tools that utilize artificial intelligence-driven predictive policing systems to forecast where crimes are likely to occur and who may commit them by analyzing large sets of data. But the expansion of predictive A.I.-systems extends beyond local police, as it was recently revealed that the FBI is now seeking to enhance its intelligence watch list systems with A.I.-predictive capabilities. The change was reported after a recent request was listed that invited vendors to bid on a project for the agency to create, “an artificial intelligence system for pre-crime policing” that would “leverage existing enterprise datasets…to develop predictive models” that “predict where additional relevant information may be derived across federated systems.” In other words, the FBI wants an artificial-intelligence powered search function to automatically sift through its reams of data to help add people to its terrorist watch list, which already numbers approximately around 2 million names, and is known to be notoriously secretive. Many individuals do not know that they are even on the lists, which numbers will undoubtedly increase once this new tool is implemented. This expansion is planned despite previous audits having found serious errors in its data in the past, with the Supreme Court even ruling against the government in multiple cases of certain individuals who challenged their inclusion on the list in the past. While originally designed for Islamic terrorists in the aftermath of 9/11, the definition of who qualifies to be on the list has greatly expanded since then. As FBI Director Patel touted on Fox News, “I’ve got every major tech company in the world embedded in the FBI, rebuilding our internet capabilities, our classified systems, and the ability for artificial intelligence to be in our counterterrorism program so we can get instantaneous results. What’s the point of collecting terabytes of data if you can’t sift through it?” But many watchdog agencies worry that the system that was already rife with potential for abuse and lack of oversight and that handing this power over to machines will only make this issue grow with the potential for real world consequences for those placed on the list.
Potential for Bias in Predictive Policing Tools
Many watchdog agencies and advocacy groups have predicted that the rise in A.I. in policing will disproportionately affect minority and other historically repressed groups. In the Department of Justice’s own report from December 2024, they warned that the data, “used for predictive policing may have significant gaps and errors, and it may reflect human biases. Use of models based on that data may entrench existing disparities and result in unintended consequences and unjust outcomes.” This can create a vicious cycle in which the aforementioned areas are subject to further over-policing that will generate further arrest data that feeds the algorithm, which then unfairly tells law enforcement to concentrate more in those areas in a sort of doom loop for residents. Not to mention such discriminatory, or in fact arbitrary, policing practices could have broad Fourth and Fifth Amendment concerns for individuals accused of crimes in those areas or whose crimes were based on the use of artificial intelligence rather than the individual judgment of a law enforcement officer. This is especially concerning because many of the private companies utilized by law enforcement lack government oversight and often rely on privately collected data which has become an increasing problem as we have previously covered extensively. Undoubtedly, just like geofence warrants, automatic license plate readers, and other mass surveillance technology, the legal challenges will continue as the surveillance state in this country expands.
